You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

372 lines
11KB

  1. /*
  2. * MP3 muxer
  3. * Copyright (c) 2003 Fabrice Bellard
  4. *
  5. * This file is part of FFmpeg.
  6. *
  7. * FFmpeg is free software; you can redistribute it and/or
  8. * modify it under the terms of the GNU Lesser General Public
  9. * License as published by the Free Software Foundation; either
  10. * version 2.1 of the License, or (at your option) any later version.
  11. *
  12. * FFmpeg is distributed in the hope that it will be useful,
  13. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  15. * Lesser General Public License for more details.
  16. *
  17. * You should have received a copy of the GNU Lesser General Public
  18. * License along with FFmpeg; if not, write to the Free Software
  19. * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  20. */
  21. #include <strings.h>
  22. #include "avformat.h"
  23. #include "avio_internal.h"
  24. #include "id3v1.h"
  25. #include "id3v2.h"
  26. #include "rawenc.h"
  27. #include "libavutil/avstring.h"
  28. #include "libavcodec/mpegaudio.h"
  29. #include "libavcodec/mpegaudiodata.h"
  30. #include "libavcodec/mpegaudiodecheader.h"
  31. #include "libavutil/intreadwrite.h"
  32. #include "libavutil/opt.h"
  33. #include "libavcodec/mpegaudio.h"
  34. #include "libavcodec/mpegaudiodata.h"
  35. #include "libavcodec/mpegaudiodecheader.h"
  36. #include "libavformat/avio_internal.h"
  37. #include "libavutil/dict.h"
  38. static int id3v1_set_string(AVFormatContext *s, const char *key,
  39. uint8_t *buf, int buf_size)
  40. {
  41. AVDictionaryEntry *tag;
  42. if ((tag = av_dict_get(s->metadata, key, NULL, 0)))
  43. av_strlcpy(buf, tag->value, buf_size);
  44. return !!tag;
  45. }
  46. static int id3v1_create_tag(AVFormatContext *s, uint8_t *buf)
  47. {
  48. AVDictionaryEntry *tag;
  49. int i, count = 0;
  50. memset(buf, 0, ID3v1_TAG_SIZE); /* fail safe */
  51. buf[0] = 'T';
  52. buf[1] = 'A';
  53. buf[2] = 'G';
  54. /* we knowingly overspecify each tag length by one byte to compensate for the mandatory null byte added by av_strlcpy */
  55. count += id3v1_set_string(s, "TIT2", buf + 3, 30 + 1); //title
  56. count += id3v1_set_string(s, "TPE1", buf + 33, 30 + 1); //author|artist
  57. count += id3v1_set_string(s, "TALB", buf + 63, 30 + 1); //album
  58. count += id3v1_set_string(s, "TDRL", buf + 93, 4 + 1); //date
  59. count += id3v1_set_string(s, "comment", buf + 97, 30 + 1);
  60. if ((tag = av_dict_get(s->metadata, "TRCK", NULL, 0))) { //track
  61. buf[125] = 0;
  62. buf[126] = atoi(tag->value);
  63. count++;
  64. }
  65. buf[127] = 0xFF; /* default to unknown genre */
  66. if ((tag = av_dict_get(s->metadata, "TCON", NULL, 0))) { //genre
  67. for(i = 0; i <= ID3v1_GENRE_MAX; i++) {
  68. if (!strcasecmp(tag->value, ff_id3v1_genre_str[i])) {
  69. buf[127] = i;
  70. count++;
  71. break;
  72. }
  73. }
  74. }
  75. return count;
  76. }
  77. #define VBR_NUM_BAGS 400
  78. #define VBR_TOC_SIZE 100
  79. typedef struct MP3Context {
  80. const AVClass *class;
  81. int id3v2_version;
  82. int write_id3v1;
  83. int64_t frames_offset;
  84. int32_t frames;
  85. int32_t size;
  86. uint32_t want;
  87. uint32_t seen;
  88. uint32_t pos;
  89. uint64_t bag[VBR_NUM_BAGS];
  90. } MP3Context;
  91. static int mp2_write_trailer(struct AVFormatContext *s)
  92. {
  93. uint8_t buf[ID3v1_TAG_SIZE];
  94. MP3Context *mp3 = s->priv_data;
  95. /* write the id3v1 tag */
  96. if (mp3 && mp3->write_id3v1 && id3v1_create_tag(s, buf) > 0) {
  97. avio_write(s->pb, buf, ID3v1_TAG_SIZE);
  98. }
  99. /* write number of frames */
  100. if (mp3 && mp3->frames_offset) {
  101. avio_seek(s->pb, mp3->frames_offset, SEEK_SET);
  102. avio_wb32(s->pb, s->streams[0]->nb_frames);
  103. avio_seek(s->pb, 0, SEEK_END);
  104. }
  105. avio_flush(s->pb);
  106. return 0;
  107. }
  108. #if CONFIG_MP2_MUXER
  109. AVOutputFormat ff_mp2_muxer = {
  110. .name = "mp2",
  111. .long_name = NULL_IF_CONFIG_SMALL("MPEG audio layer 2"),
  112. .mime_type = "audio/x-mpeg",
  113. .extensions = "mp2,m2a",
  114. .audio_codec = CODEC_ID_MP2,
  115. .video_codec = CODEC_ID_NONE,
  116. .write_packet = ff_raw_write_packet,
  117. .write_trailer = mp2_write_trailer,
  118. .flags = AVFMT_NOTIMESTAMPS,
  119. };
  120. #endif
  121. #if CONFIG_MP3_MUXER
  122. static const AVOption options[] = {
  123. { "id3v2_version", "Select ID3v2 version to write. Currently 3 and 4 are supported.",
  124. offsetof(MP3Context, id3v2_version), AV_OPT_TYPE_INT, {.dbl = 4}, 3, 4, AV_OPT_FLAG_ENCODING_PARAM},
  125. { "write_id3v1", "Enable ID3v1 writing. ID3v1 tags are written in UTF-8 which may not be supported by most software.",
  126. offsetof(MP3Context, write_id3v1), AV_OPT_TYPE_INT, {.dbl = 0}, 0, 1, AV_OPT_FLAG_ENCODING_PARAM},
  127. { NULL },
  128. };
  129. static const AVClass mp3_muxer_class = {
  130. .class_name = "MP3 muxer",
  131. .item_name = av_default_item_name,
  132. .option = options,
  133. .version = LIBAVUTIL_VERSION_INT,
  134. };
  135. static const int64_t xing_offtbl[2][2] = {{32, 17}, {17,9}};
  136. /*
  137. * Write an empty XING header and initialize respective data.
  138. */
  139. static int mp3_write_xing(AVFormatContext *s)
  140. {
  141. AVCodecContext *codec = s->streams[0]->codec;
  142. MP3Context *mp3 = s->priv_data;
  143. int bitrate_idx = 3;
  144. int64_t xing_offset;
  145. int32_t mask, header;
  146. MPADecodeHeader c;
  147. int srate_idx, i, channels;
  148. int needed;
  149. for (i = 0; i < FF_ARRAY_ELEMS(avpriv_mpa_freq_tab); i++)
  150. if (avpriv_mpa_freq_tab[i] == codec->sample_rate) {
  151. srate_idx = i;
  152. break;
  153. }
  154. if (i == FF_ARRAY_ELEMS(avpriv_mpa_freq_tab)) {
  155. av_log(s, AV_LOG_ERROR, "Unsupported sample rate.\n");
  156. return -1;
  157. }
  158. switch (codec->channels) {
  159. case 1: channels = MPA_MONO; break;
  160. case 2: channels = MPA_STEREO; break;
  161. default: av_log(s, AV_LOG_ERROR, "Unsupported number of channels.\n"); return -1;
  162. }
  163. /* dummy MPEG audio header */
  164. header = 0xff << 24; // sync
  165. header |= (0x7 << 5 | 0x3 << 3 | 0x1 << 1 | 0x1) << 16; // sync/mpeg-1/layer 3/no crc*/
  166. header |= (srate_idx << 2) << 8;
  167. header |= channels << 6;
  168. for (;;) {
  169. if (15 == bitrate_idx)
  170. return -1;
  171. mask = (bitrate_idx << 4) << 8;
  172. header |= mask;
  173. avpriv_mpegaudio_decode_header(&c, header);
  174. xing_offset=xing_offtbl[c.lsf == 1][c.nb_channels == 1];
  175. needed = 4 // header
  176. + xing_offset
  177. + 4 // xing tag
  178. + 4 // frames/size/toc flags
  179. + 4 // frames
  180. + 4 // size
  181. + VBR_TOC_SIZE; // toc
  182. if (needed <= c.frame_size)
  183. break;
  184. header &= ~mask;
  185. ++bitrate_idx;
  186. }
  187. avio_wb32(s->pb, header);
  188. ffio_fill(s->pb, 0, xing_offset);
  189. avio_wb32(s->pb, MKBETAG('X', 'i', 'n', 'g'));
  190. avio_wb32(s->pb, 0x01 | 0x02 | 0x04); // frames/size/toc
  191. mp3->frames_offset = avio_tell(s->pb);
  192. mp3->size = c.frame_size;
  193. mp3->want=1;
  194. mp3->seen=0;
  195. mp3->pos=0;
  196. avio_wb32(s->pb, 0); // frames
  197. avio_wb32(s->pb, 0); // size
  198. // toc
  199. for (i = 0; i < VBR_TOC_SIZE; ++i)
  200. avio_w8(s->pb, (uint8_t)(255 * i / VBR_TOC_SIZE));
  201. ffio_fill(s->pb, 0, c.frame_size - needed);
  202. avio_flush(s->pb);
  203. return 0;
  204. }
  205. /*
  206. * Add a frame to XING data.
  207. * Following lame's "VbrTag.c".
  208. */
  209. static void mp3_xing_add_frame(AVFormatContext *s, AVPacket *pkt)
  210. {
  211. MP3Context *mp3 = s->priv_data;
  212. int i;
  213. ++mp3->frames;
  214. mp3->size += pkt->size;
  215. if (mp3->want == ++mp3->seen) {
  216. mp3->bag[mp3->pos] = mp3->size;
  217. if (VBR_NUM_BAGS == ++mp3->pos) {
  218. /* shrink table to half size by throwing away each second bag. */
  219. for (i = 1; i < VBR_NUM_BAGS; i += 2)
  220. mp3->bag[i >> 1] = mp3->bag[i];
  221. /* double wanted amount per bag. */
  222. mp3->want <<= 1;
  223. /* adjust current position to half of table size. */
  224. mp3->pos >>= 1;
  225. }
  226. mp3->seen = 0;
  227. }
  228. }
  229. static void mp3_fix_xing(AVFormatContext *s)
  230. {
  231. MP3Context *mp3 = s->priv_data;
  232. int i;
  233. avio_flush(s->pb);
  234. avio_seek(s->pb, mp3->frames_offset, SEEK_SET);
  235. avio_wb32(s->pb, mp3->frames);
  236. avio_wb32(s->pb, mp3->size);
  237. avio_w8(s->pb, 0); // first toc entry has to be zero.
  238. for (i = 1; i < VBR_TOC_SIZE; ++i) {
  239. int j = i * mp3->pos / VBR_TOC_SIZE;
  240. int seek_point = 256LL * mp3->bag[j] / mp3->size;
  241. avio_w8(s->pb, FFMIN(seek_point, 255));
  242. }
  243. avio_flush(s->pb);
  244. avio_seek(s->pb, 0, SEEK_END);
  245. }
  246. /**
  247. * Write an ID3v2 header at beginning of stream
  248. */
  249. static int mp3_write_header(struct AVFormatContext *s)
  250. {
  251. MP3Context *mp3 = s->priv_data;
  252. int ret;
  253. ret = ff_id3v2_write(s, mp3->id3v2_version, ID3v2_DEFAULT_MAGIC);
  254. if (ret < 0)
  255. return ret;
  256. if (s->pb->seekable)
  257. mp3_write_xing(s);
  258. return 0;
  259. }
  260. static int mp3_write_packet(AVFormatContext *s, AVPacket *pkt)
  261. {
  262. if (! pkt || ! pkt->data || pkt->size < 4)
  263. return ff_raw_write_packet(s, pkt);
  264. else {
  265. MP3Context *mp3 = s->priv_data;
  266. #ifdef FILTER_VBR_HEADERS
  267. MPADecodeHeader c;
  268. int base;
  269. ff_mpegaudio_decode_header(&c, AV_RB32(pkt->data));
  270. /* filter out XING and INFO headers. */
  271. base = 4 + xing_offtbl[c.lsf == 1][c.nb_channels == 1];
  272. if (base + 4 <= pkt->size) {
  273. uint32_t v = AV_RB32(pkt->data + base);
  274. if (MKBETAG('X','i','n','g') == v || MKBETAG('I','n','f','o') == v)
  275. return 0;
  276. }
  277. /* filter out VBRI headers. */
  278. base = 4 + 32;
  279. if (base + 4 <= pkt->size && MKBETAG('V','B','R','I') == AV_RB32(pkt->data + base))
  280. return 0;
  281. #endif
  282. if (mp3->frames_offset)
  283. mp3_xing_add_frame(s, pkt);
  284. return ff_raw_write_packet(s, pkt);
  285. }
  286. }
  287. static int mp3_write_trailer(AVFormatContext *s)
  288. {
  289. MP3Context *mp3 = s->priv_data;
  290. int ret=mp2_write_trailer(s);
  291. if (ret < 0)
  292. return ret;
  293. if (mp3->frames_offset)
  294. mp3_fix_xing(s);
  295. return 0;
  296. }
  297. AVOutputFormat ff_mp3_muxer = {
  298. .name = "mp3",
  299. .long_name = NULL_IF_CONFIG_SMALL("MPEG audio layer 3"),
  300. .mime_type = "audio/x-mpeg",
  301. .extensions = "mp3",
  302. .priv_data_size = sizeof(MP3Context),
  303. .audio_codec = CODEC_ID_MP3,
  304. .video_codec = CODEC_ID_NONE,
  305. .write_header = mp3_write_header,
  306. .write_packet = mp3_write_packet,
  307. .write_trailer = mp3_write_trailer,
  308. .flags = AVFMT_NOTIMESTAMPS,
  309. .priv_class = &mp3_muxer_class,
  310. };
  311. #endif